Spatiotemporal-frequency analysis applied to motion detection

Cardiac wall motion is related to the health of the heart. Left ventricular contractile abnormalities can be an important manifestation of coronary artery disease. Wall motion changes may represent ischemia or infarction of myocardium, which makes the importance of this research problem. In this work we describe a method for motion detection based on a spatiotemporal-frequency (STF) analysis which computes the optical flow field from a series of gated-SPECT images. Wigner-Ville distribution (WVD) is used in STF analysis to obtain the signal frequency occurrence in time and space. By the STF approach the motion estimation requires the determination of the instantaneous frequency which is obtained from the first moment or by peak detection of the WVD. Frequency constraints are used to formulate equations to obtain the velocity vector field within a single image neighborhood. Preliminary results have shown the feasibility of the proposed methodology and the application in gated-SPECT images in myocardial perfusion studies is currently being validated.